The Menu Profit Fix

The problem

You sit down in the back office on a Tuesday afternoon and pull up the P&L. Food cost is 31%. Right where the trade magazines say it should be. Saturday was full, Friday was a wait, catering had a good month. And yet the bank account looks the same as March and the check you wrote yourself this month is the smallest one in two years. Where did the money go?

Most of the time it went out one dollar at a time on a dish you have not repriced since 2023. The carbonara that sells 500 plates a month at $18 when the market would absorb $20 is leaking $1,000 a month. The build-your-own omelette stuck at $11.50 since the egg shortage of 2021 is bleeding $2,000. The cheese pizza at $14 with a $5.60 contribution margin instead of the $8 your menu average demands. Each one feels too small to bother with. Each one is selling. The food cost percentage on each one looks fine. Together they are the structural reason your net margin is 6% instead of 12%.

The average US independent restaurant runs on a 3-9% net margin. Sixty percent of independents close within five years. The gap between thriving and closing is, on most menus I have walked through, eight to fifteen mispriced dishes leaking $500 to $1,500 a month each. The food cost percentage column does not see them. The contribution margin column does.

What most people get wrong

They chase food cost percentage as if it were a goal instead of a diagnostic. The food cost line on the P&L is the number every owner stares at because it is what the POS calculates automatically and what the trade press obsessed over for forty years. An operator sees a 28% food-cost dish next to a 35% food-cost dish and assumes the 28% is the better dish. It is almost never true. A coffee shop runs 20% food cost and goes bankrupt because labor is 45%. A steakhouse runs 38% food cost and prints money because each ticket is $80. The percentage is a ratio. Your rent is denominated in dollars. The metric that pays rent is contribution margin: sell price minus total plate cost, in actual currency. The book is structured around this mental shift and the math is so direct that operators usually grasp it in the first chapter and start rewriting their cost cards before they finish the second.

They reprice by rounding up to the next “nice” number and stop. The carbonara went from $16 to $17 in 2024. The omelette went from $10.99 to $11.50 in 2022. The owner felt the increase, customers did not flinch, and the menu sat unchanged for two years while supplier costs marched. There is no governance, no trigger, no schedule. The book replaces “round up and hope” with a structured 30-day sprint that uses CM data to identify which specific dishes to touch, by how much, in what sequence: cut dogs in week two, reprice workhorses in week three, reposition puzzles in week four. The dollar recovery on a $50,000-a-month operation is typically $1,500-$3,000 in the first sprint and the discipline compounds across quarterly resets.

They forget that DoorDash and Uber Eats are taxes that did not exist when the menu was priced. A burger priced at $14 with a $5 plate cost contributes $9 to your business when a guest sits down. The same burger on DoorDash, after the 30% commission and $0.75 in packaging the app does not reimburse, contributes $4.05. The owner sees the same $14 menu price and assumes the same $9 margin. The delivery app pulled $4.95 out without anyone noticing. At a quarter of orders going through the apps, that is $745 a month in invisible margin compression on a single dish. The fix is a three-line spreadsheet that computes blended CM by channel and either reprices the app menu, drops a tier, or quits an app that does not earn its keep.

A working approach

The book is built around five tools you assemble in sequence. Each one stands alone and each one feeds the next. There is no software requirement. Excel and a kitchen scale do the work.

STAGE 1 — Recipe cost card (per dish)
  Ingredient * weight * per-unit cost = plate cost
  + Ancillary line (10% baseline) = total plate cost

STAGE 2 — Contribution margin worksheet
  Sell price - Total plate cost = CM per plate
  CM per plate * Units sold (90d) = Monthly CM
  Sort by Monthly CM, descending

STAGE 3 — Menu matrix (BCG-derived)
  CM (vertical) x Menu mix % (horizontal) =
  Stars / Workhorses / Puzzles / Dogs

STAGE 4 — Supplier price tracker
  4-7 ingredients drive 70% of food spend
  Track weekly, trigger at 6-12% movement

STAGE 5 — 30-day reprice sprint
  Week 1: Measure   |  Week 2: Cut dogs
  Week 3: Reprice   |  Week 4: Reposition

The recipe cost card is the foundational tool, and it is where most independents stop pretending they know their numbers. The honest version captures two things the casual version misses. First, cooked yield loss: a 6-oz raw chicken breast cooks down to 4.5 oz, so a recipe calling for “6 oz grilled chicken” actually starts with 8 oz raw and the cost card was understating protein cost by a third. Second, ancillary cost: the lemon wedge, the to-go clamshell on 30% of orders, the napkin, the disposable cutlery, the staff meal, the modifications absorbed because the line cook never charges for almond milk. Most operators run an 8-18% gap between their headline plate cost and their honest total plate cost. On a dish you thought cost $5, the real number is $5.80-$5.90. Your real food cost is 37%, not 31%. The first time you build the honest cost card on a top-five dish, you find money sitting in plain sight.

CM math vs food-cost percentage

The chapter-three example flips the lights on. Two dishes side by side: a house salad at $8 with $1.40 food cost (17.5%) selling 80 plates a month, and a 12-oz ribeye at $38 with $13.30 food cost (35%) selling 220 plates a month. Food cost percentage says the salad is the better dish. Contribution margin says something different: the salad puts $528 a month into the business, the ribeye puts $5,434. The dish with the 35% food cost generates ten times more dollars than the dish with the 17.5%. The percentage was lying. Print this and tape it to the office wall: rent is paid in dollars, not percentage points. The number you run on every dish is CM, the number you rank your menu by is monthly CM, and the dish you protect most jealously is whichever sits at the top of that ranked list, regardless of food cost percentage.

The CM worksheet is eight columns: dish name, total plate cost, menu price, CM per plate, units sold last 30 days, monthly CM, menu mix percent, CM rank. Build it once. Update monthly. Run a finger down the list looking for dishes whose CM sits below your menu average. Those are your action items. Healthy independent full-service spots aim for an average CM per plate of $8.50 or higher, with a CM floor of $5-7 below which the dish has to be repriced, reformulated, or removed.

The BCG menu matrix

The two-by-two matrix is the chapter-four framework that turns the CM worksheet into a one-page action list. Contribution margin on the vertical axis, menu mix percentage on the horizontal. Four quadrants, four actions.

HIGH CM
  |
  |   PUZZLE         STAR
  |   (reposition)   (protect)
  |
  +---------------------------- MIX %
  |
  |   DOG            WORKHORSE
  |   (cut)          (reprice)
  |
LOW CM

A star contributes above your menu average per plate and sells above your average mix. The 12-oz ribeye is a star. So is the Buffalo wing platter when the wing price has not spiked. The action is protect: re-cost the recipe every quarter so portion drift does not eat the margin you built, spot-check the line for portion creep, never substitute suppliers without re-costing first. You can reprice a star, but only 4-7% per move, never more than twice a year, and only when the trailing-90-day mix is stable. The expensive mistake with stars is overconfidence: “everyone loves it, they will pay more.” Often true, up to a price point you cannot identify from inside the restaurant.

A workhorse sells well and contributes below average. The carbonara from this article’s opening is a workhorse. The cheese pizza at $14. The half-pound burger stuck at $13 in a market that would now absorb $16. Workhorses are the most dangerous category because volume hides the margin problem. The action is almost always reprice: a $14 dish moves to $16, volume drops 8-10%, CM jumps from $5 to $7, monthly contribution goes from $2,000 to $2,520, you added $520 with one decision. When the market will not bear the reprice because the dish is anchored to a competitor’s identical product, you reformulate instead: shave 0.5 oz off the portion, swap premium oil to a cooking blend, move a $0.70 garnish to a $0.30 garnish that plates the same. A skillful 8% cost reduction on a workhorse equals a 5% price increase in net margin impact and the customer never sees the change.

A puzzle is a high-CM dish that does not sell enough. The truffle risotto. The pan-seared salmon buried mid-section with the description “salmon, vegetables, rice.” The recipe is healthy. You have a distribution problem. The action is reposition: move it into the upper-right of the menu page (the “golden triangle” where eye-tracking shows guests look first), rewrite with provenance plus method plus sensory detail (“wild Alaskan king salmon over saffron risotto with charred Meyer lemon” instead of “grilled salmon”), train servers to recommend it by name. A puzzle that moves into the star quadrant is the single highest-leverage decision on the menu because you do not add a dish or do new prep, you unlock margin that was already sitting on the page. The Brooklyn cafe case study rewrote 14 descriptions on a Tuesday afternoon, made no other changes, and lifted blended ticket size 8.4% over six weeks. About $3,100 a month for one afternoon of writing.

A dog is a low-CM dish that does not sell. There is no reason for it to be on your menu. You are paying inventory cost, training cost, prep waste, and menu real estate for a dish that does neither margin nor volume. The action is cut. Two narrow exceptions: a strategic gateway dish that anchors a price tier, or a dish using an ingredient critical to multiple other dishes where removing it forces a supply change. Both are real, both are uncommon. The suburban Chicago Italian spot in chapter four cut nine dogs and consolidated four ambiguous items into two; six months later monthly CM was up 11%, prep waste was down 22%, line-cook throughput was up. The owner’s quote: “the menu felt scary-short the first week. By month two we wondered why we ever had the dogs in the first place.”

Supplier price triggers (not annual reprice)

Most restaurants reprice once or twice a year. Most supplier prices change every two to four weeks. The gap is where margin disappears. When wing prices jump 18% in March and your menu does not move until November, you have absorbed eight months of margin compression on a high-volume item. The fix is a tracker plus a trigger: track the few ingredients that drive the math, set a threshold at which you act, and act when the trigger fires.

Four to seven ingredients account for 70% of total food spend on a typical independent menu. Full-service: primary proteins (beef, chicken, salmon), cooking oils, cheese, dairy, produce staples (tomatoes, avocados, romaine). Burger spot: beef, cheese, buns, potatoes. Pizza shop: flour, cheese, tomato, pepperoni. Identify your own list by pulling 90 days of supplier invoices, sorting by total spend, taking the items that cumulatively represent 70-75% of food spend. The tracker is eight columns in a weekly spreadsheet: ingredient, supplier, unit, last price, current price, change percent, trigger threshold, action. Five minutes a week. Trigger threshold scales with how much the ingredient drives your menu: anchor ingredients (15%+ of any dish’s cost) trigger at 6%, heavy ingredients (5-15% of dish cost) at 10%, standard ingredients (1-5%) at 15%, trace items you do not track at all.

When a row triggers you have three responses. First, source differently: a new supplier, a different cut, a different brand. Get three competing quotes before you do anything customer-facing, because six out of ten triggers resolve with a phone call. Second, adjust the portion: 6 oz becomes 5.5 oz, 2 oz of sauce becomes 1.75 oz, small portion adjustments rarely register if the plate still looks generous. A 10% portion reduction offsets a 10% cost increase with zero menu reprice. Third, when sourcing and portion both fail, reprice the affected dishes. If beef went up $1.20 per pound and your burger uses 8 oz, the dish cost moved by $0.60, so a $1 menu price increase covers it with a cushion. Reprice quietly. The Austin barbecue spot caught a 14% brisket cost jump in the first week of March, made three moves the same week (partial supplier swap, 1 oz portion trim, $2 reprice on the brisket plate), and grew monthly CM by $1,800 the next full month.

DoorDash, Uber Eats, Grubhub margin math

Delivery apps did not exist as a meaningful share of independent revenue in 2018. By 2026 they are 18-35% of revenue at the average independent full-service spot, and almost no menu has been repriced to reflect the change. DoorDash takes 15-30% commission depending on tier (Basic, Plus, Premier). Uber Eats takes 15-30% on marketplace plus tiers. Grubhub takes 10-30%, with the cheapest tier reserved for restaurants handling their own delivery. Each is a top-line percentage off your gross sale, applied before food cost is even paid. The 30% DoorDash Plus commission is bigger than the food cost on most independent menus.

The worked example is a burger priced at $14 with a $5 plate cost. Dine-in CM is $9 (64%). On DoorDash: $14 menu price, minus $4.20 commission (30%), minus $5 plate cost, minus $0.75 packaging the app does not reimburse, CM falls to $4.05 (29%). Same dish, same kitchen, blended CM down 55%. If a quarter of your burger orders go through DoorDash, your blended CM is not $9, it is $7.76. On 600 plates a month that is $745 in monthly margin the app pulled out without anyone noticing. The book walks through three strategies. Pricing parity (charge the same on the app as in-house) is the default and is the structural reason most independent margins compressed through 2020-2024. Delivery pricing premium (15-20% more on the app menu) explicitly offsets the commission and is now industry standard since DoorDash legitimized in-app pricing. A delivery-specific menu (smaller, repriced for app economics, only dishes that survive 20 minutes in a clamshell) is worth the maintenance if delivery is more than 25% of revenue. The highest-leverage move is shifting orders from the marketplace apps to direct ordering through Toast Online, Square Online, or ChowNow, where a 12% discount on a $30 direct order costs $3.60 versus the $9 the same order would have cost on DoorDash. The discount is the cheaper version of the same customer.

The 30-day reprice sprint

The implementation chapter ties everything into a four-week sequence. Most operators try to fix everything at once: they reprice 20 dishes in a single menu update, regulars notice, social media lights up, and they reverse course within six weeks. The reputation damage is real, the margin recovery is incomplete, and the experience teaches them not to try again for years. The sprint avoids this by sequencing the changes: cut the dogs first, reprice the workhorses next, reposition the puzzles after. Changes are spread across menu sections so no single customer reads three price increases on the same page. Changes are not announced. By the time anyone notices, the trailing-30-day CM is up $2,000 a month and the question of whether it worked has already been answered by the bank.

Week one is preparation. No price changes. Pull 90-day item-level sales from the POS (Toast, Square, Clover, Lightspeed, TouchBistro all export cleanly), build the recipe cost cards, run the CM worksheet. End of week one: a one-page action sheet with dishes to cut, reprice, and reposition. Week two is cuts. 86 the dogs in the POS, train servers on the substitute script (“we retired that one, can I tell you about [substitute]”), reprint the menu, update delivery apps. Expect one or two angry-customer moments and a small comp resolves each one. Week three is the dollar lever. Reprice workhorses 4-8% each, round to natural numbers ($14 or $15, never $14.32), spread across sections, update all menu surfaces quietly, watch the next four days for unusual mix shifts. A 5-12% volume drop is normal. A 25%+ drop suggests you overshot and you pull back. Week four is the upside. Rewrite puzzle descriptions in the provenance-method-sensory pattern, reposition them in the golden triangle, train servers with a one-page recommendation card.

The case study is a 28-seat family Italian restaurant in central New Jersey. Pre-sprint: $52,000 monthly sales, 31% food cost, $3,400 net profit (6.5% margin). Week one math identified seven dogs, five workhorses, four puzzles. Week two cut the seven dogs. Week three repriced the five workhorses by an average $1.40 per dish, with volume dropping about 7%. Week four rewrote four puzzle descriptions and repositioned two of them. Post-sprint: $53,500 monthly sales (up 3%), 30.5% food cost, $6,200 net profit (11.6% margin). About nine hours of work over 30 days. $2,800 a month in recovered CM. The owner ran a second sprint at the 90-day mark and the third sprint took about six hours because the templates were already built. The cadence compounds: every 90 days a fresh sprint keeps the menu ahead of cost drift and prevents the slow margin compression that closes most independents over a decade of inattention.

Where this scales

The article walked through the five tools. The book covers each with worked Excel templates and case studies from operators who have run versions of the sprint with me. Two CSV templates ship in the bonus folder: menu-engineering-worksheet.csv has the contribution margin structure pre-populated with sample dishes (carbonara, ribeye, omelette, cheese pizza) so you can replace each row with your own data and see your menu sorted honestly for the first time. supplier-price-tracker.csv has the eight-column tracking structure with thirty representative ingredients across proteins, dairy, oils, and produce so you can drop your own invoice prices in and watch the trigger column flag movement.

The book is opinionated about software. You do not need MarketMan at $149 a month or Restaurant365 at $435 a month to run this system. Excel does the work for 95% of independents. The discipline matters more than the tool: I have walked into restaurants paying $5,200 a year for Restaurant365 with recipe cards last updated 18 months ago. Toast Inventory at $75 a month or Square for Restaurants (included in the POS plan) are sensible upgrades if you already use those POS systems and want recipe costing tied to invoices automatically. Anything more sophisticated is justified only once you are running multiple locations or have crossed about $1.5 million in annual revenue. Wholesale sourcing helps too: profit.deals carries case-pack pricing on the ingredient categories that drive the trigger column (proteins, oils, cheese, dairy, produce staples) at restaurant supply tiers an independent buyer normally cannot access alone.

Included with the book

• Menu Engineering Worksheet (CSV) — the contribution margin structure pre-populated with sample dishes. Drop your own data in, sort by Monthly CM, classify into stars/workhorses/puzzles/dogs.
• Supplier Price Tracker (CSV) — eight columns covering ingredient, supplier, unit, last price, current price, change percent, trigger threshold, and action. Thirty representative ingredients pre-populated. Update weekly with new invoices.
• 30-Day Reprice Sprint Checklist (Markdown) — the chapter eight sequence broken into a week-by-week task list you can print and tape to the office wall.

Get the full picture

Readers of this also chose

Questions readers ask

Does this work for QSR and fast-casual or only full-service?

The contribution margin math is the same across formats; the food cost targets and CM floors differ. Full-service aims for 28-35% food cost and $8.50+ average CM. QSR and fast-casual aim for 25-30% food cost with a lower per-plate CM and higher volume. The matrix, the supplier tracker, and the 30-day sprint all work identically.

What about bars and beverage programs?

The same framework. Bar program pour cost target is 18-22%. The CM worksheet, supplier tracker, and reprice sprint apply directly to wine, beer, and cocktails. The book includes beverage examples in the worked CM tables (Espresso Martini, House Negroni, House Red by glass and bottle).

What if I need a refund?

Checkout runs on Lemon Squeezy. The standard refund window applies. You keep the PDF and the CSV templates either way.

How long until I see the margin recovery in the bank?

The 30-day sprint produces visible CM lift in the trailing 30-day numbers immediately. The cash conversion (timing of supplier invoices, payroll cycles) means the bank account reflects it about 45-60 days after the sprint completes. The compounding effect across quarterly sprints is where the multi-thousand-dollar monthly recovery shows up.